1. Field of the Invention
The present invention relates to a high frequency compensator for compensating high frequency component of a compressed digital audio signal such as a mini disk (MD), a moving picture experts group 1 audio layer 3 (MP3), and a digital audio broadcasting (DAB), and relates to a reproducing device includes the high frequency compensator and reproduces a digital audio signal.
2. Description of the Related Art
With respect to a compact disc (CD), an audio signal is a stereo signal having the sampling frequency of 44.1 kHz and the data bit length of 16 bits. A data rate can be expressed by the number of bits required for 1 second. The data rate of the CD is 1411.2 kbps. With respect to the MD, the data volume to be encoded is reduced by advanced transform acoustic coding (ATRAC) and ATRAC 3 which is an improved version of the ATRAC. The ATRAC and the ATRAC 3 are compression technologies for decimating a component that a human being can hardly hear and the data before and after big sound by utilizing the characteristic of human auditory sense. The ATRAC 3 is capable of the data compression rates of two times (mini disk long play mode (MDLP) 2, 132 kbps) and four times (MDLP 4, 66 kbps) those of the ATRAC. Therefore, the ATRAC 3 can compress an amount of data up to 1/20 level of the CD. However, as the compression rate becomes higher, the sound quality becomes deteriorated.
It is said that the audible frequency band of humans is from 20 Hz to 20 kHz. However, the audible frequency band varies between individuals, and it is considered that the upper limit for ordinal people is between 14 kHz and 15 kHz. Compression technologies have been developed in consideration of such audible frequency band, and pure sound (sine wave) is used for measuring the audible frequency band. However, in a real music or natural sound, sounds of a plurality of frequencies exist in mixed state. Therefore, there has been a widespreading idea that although a sound outside of the audible frequency band cannot be heard alone, it affects human audibility in terms of soundscape, or the like.
In Japanese Patent No. 3576941 and Japanese Patent No. 3576951, there is disclosed a frequency decimating device in which one of a pair of strongly correlated bands is removed, and in which the spectrum in a range of the other band is shortened to be onto a low frequency side, based on a spectrum distribution of a signal to be subjected to the decimating processing.
In Japanese Patent No. 3576942, there is disclosed a frequency compensating system in which when a signal is compensated, a portion having a higher spectrum distribution correlation of the spectrum of the signal to be compensated, is added along an envelope onto the high frequency side of the signal to be compensated, thereby expanding the band.
In Japanese Patent Laid-open No. 2004-80635, there is disclosed a signal encoding device which the spectrum in a limited band is encoded, the limited band corresponding to a predetermined frequency band of the input time series signal, and on the decoding side, in order to obtain the time series signal in the frequency band to be expanded based on a spectrum map, mapping information showing a mapping method is adaptively created, and the encoded spectrum in the limited band and the mapping information are output.
When a digital audio signal is compressed with a high compression rate and recorded, for example, in the MDLP 4, the frequency spectrum is only present up to 12 kHz level, as shown in FIG. 1B. Therefore, when it is compared with the CD signal shown in FIG. 1A, high frequency information is lost.